Hacker News Reader

Comment by miah_ | Hacker News Reader

Its supercharging the worst in everything its applied to. Try getting a job right now. All of the automated systems that are doing screening will throw you out if you aren't 100% compatible with whatever they're looking for. Sure, its cheaper than human screening, but now you've got a position not filled for even longer, which is probably costing you in other ways.

View on HN

> screening will throw you out if you aren't 100% compatible with whatever they're looking for

how does one know if this it the reason they are being thrown out.

You can never precisely be sure, but here are some of the behaviors I’ve been observing that give me the feeling I was auto-rejected:

- rejection coming in nearly instantly - rejection coming at “odd” times — weekend evenings, 3 AM in my time zone for a company in my time zone

The latter one doesn’t hold up too well for remote-first companies but I would hope few tech firms have recruiting staff going through applications on weekends.

Yup. I applied for a PM position, within 20 minutes of it going up. Quite a specific one: "'expert level' health insurance system knowledge and experience integrating between multiple EHR vendors and partners'". I have all of this. Have built claims benefits management systems being used by some household names in healthcare. I'm an expert in EPIC, Cerner, ESOsuite EHR systems.

10 minutes later: "we are looking for someone whose skillset and experience better aligns with our requirements and the position".

I guess my resume needed more buzzwords.

You make up a CV that exactly matches the job advert and see whether there's a response

This doesn't rule out that there were enough good or "better" applicants that crowded yours out.

Of course it does, you send a few of these perfect CVs and see what happens.

Why would a human also not choose a perfect CV over other options. What does this prove about AI.

You want a perfect candidate not a perfect CV. A perfect CV is highly likely to have lies and was made up based on the job ad, addressing every single requirement bullet.

If I sent out completely truthful CVs, I'd still be working a $15/hr data entry job or whatever.

Obviously you can't fake it forever if you don't have the skills to back it up, but the reason people lie and embellish on CVs is because they know they are reasonably fast learners, so any skills mismatch can be quickly adjusted while on the job. Easier to ask for forgiveness than permission, etc. etc.

This pretty much became a necessity since ~2005, when IT demand skyrocketed and became mainstream and automatic candidate rejection and filtering became pervasive

Even without the lies, you can meet every single bullet point and still be a worse fit than someone who doesn't match every single bullet point.

Maybe the gal or guy who fits every bullet point has a long career and quite broad knowledge. Maybe the gal or guy who doesn't fit every single bullet point is much smarter, a much harder worker with far more grit, and a quick learner.

You end up hiring the first person who may be mediocre, and missing out on the second who may be a super high performer.

> A perfect CV is highly likely to have lies and was made up based on the job ad, addressing every single requirement bullet.

Wouldn't "AI" also know this though. I would imagine AI is doing more than keyword match type stuff.

Human can look at a CV which doesn't match but say I like what this person has done, learning our stack should be no big deal.

As a hiring manager, based on the average candidate that gets forwarded to me after presumably several rounds of filtering (some mechanical, some human), the idea that there is some kind of smart matching going on is slightly ridiculous: they can't tell a security-focused software developer resume from an assurance-focused security specialist, for example.

I guess it hurts less if you can blame ai?

AI isn't scanning job applications lol ... lazy recruiters can't review them fast enough.

Reply on news.ycombinator.com